hghooks: hook to trigger cache population (
bug 1358239); r?glob
Mozilla has historically had problems with the tags cache causing
performance problems. Most of these problems were resolved a few years
ago by a rewritten tags cache implementation in Mercurial. However,
for that cache to work it needs to be populated. And, the cache isn't
populated until something attempts to resolve a symbol that could be
a tag.
This is normally not a problem. On hgweb, the tags cache will be
updated on most page views. However, it can be problematic on the
master server.
Nothing in the regular push flow appears to access tags data. This
means that the tags cache may not be written on the master server.
It so happens that we have processes running on the master server
that don't have write privileges to repos. So while they may trigger
tags cache population, the cache I/O fails and the cache is never
written. Assuming the cache is never written, over time these processes
have to compute more and more tags data and over time the amount of CPU
required balloons. Enough time passes and alerts start to fire because
processes that should be quick are spending dozens of seconds computing
tags data.
This commit solves that problem by implementing a pre transaction close
hook that accesses tags data, thus ensuring the tags cache is up to
date. Other processes should never have to compute tags data again.
Furthermore, `hg pull` will transfer tags cache data. So populating
the cache before replication ensures that hgweb machines also don't
have to populate the cache. Finally, the bundles generated for
clone bundles should have current tags caches, ensuring that people
who clone from them don't need to regenerate the data. So ensuring
the tags cache is populated on the master server is full of wins.
The new hook is globally installed on the master server so all repos
benefit from tags cache generation.
Unless we pre-populate the tags cache on all repos, the first push to
a repo after this is enabled may be slow.
All subsequent pushes may slow down because of this hook. The tags
cache population time is proportional to the number of files in a
repo and the number of heads being pushed. However, the common case
is 1 head per push and Mercurial should cache the manifest for that
head, thus assuring rapid tags cache generation. The tags cache
generation time is recorded in blackbox logs. So if the logs show
this hook causes too much of a perf drain, we can look into alternate
mechanisms for populating the tags cache.
MozReview-Commit-ID: 5q9TuqO4JaS
--- a/ansible/roles/hg-ssh/templates/hgrc.j2
+++ b/ansible/roles/hg-ssh/templates/hgrc.j2
@@ -17,16 +17,18 @@ changegroup.a_recordlogs = /var/hg/versi
changegroup.push_printurls = python:mozhghooks.push_printurls.hook
changegroup.z_advertize_upgrade = python:mozhghooks.advertise_upgrade.hook
#pretxnchangegroup.renamecase = python:mozhghooks.prevent_case_only_renames.hook
# Disabled because too many people are running into issues. Need more
# granular checking for now. Bug 787620.
#pretxnchangegroup.author_format = python:mozhghooks.author_format.hook
pretxnchangegroup.single_root = python:mozhghooks.single_root.hook
+pretxnclose.populate_caches = python:mozhghooks.populate_caches.hook
+
[extensions]
blackbox =
clonebundles =
obsolescencehacks = /var/hg/version-control-tools/hgext/obsolescencehacks
pushlog = /var/hg/version-control-tools/hgext/pushlog
serverlog = /var/hg/version-control-tools/hgext/serverlog
readonly = /var/hg/version-control-tools/hgext/readonly
new file mode 100644
--- /dev/null
+++ b/hghooks/mozhghooks/populate_caches.py
@@ -0,0 +1,10 @@
+# This software may be used and distributed according to the terms of the
+# GNU General Public License version 2 or any later version.
+
+from __future__ import absolute_import
+
+
+def hook(ui, repo, **kwargs):
+ # Trigger tags cache generation.
+ repo.tags()
+ repo.unfiltered().tags()
new file mode 100644
--- /dev/null
+++ b/hghooks/tests/test-populate-caches.t
@@ -0,0 +1,65 @@
+ $ hg init server
+ $ cat >> server/.hg/hgrc << EOF
+ > [extensions]
+ > blackbox =
+ > [blackbox]
+ > track = *
+ > EOF
+
+ $ hg -q clone server client
+ $ cd client
+
+ $ touch foo
+ $ hg -q commit -A -m initial
+ $ hg push
+ pushing to $TESTTMP/server
+ searching for changes
+ adding changesets
+ adding manifests
+ adding file changes
+ added 1 changesets with 1 changes to 1 files
+
+No tags cache should exist because there is no .hgtags file
+
+ $ [ -f ../server/.hg/cache/hgtagsfnodes1 ]
+ [1]
+
+Pushing a tag should not populate the tags cache unless without the hook
+
+ $ hg tag initial
+ $ hg -q push
+
+ $ [ -f ../server/.hg/cache/hgtagsfnodes1 ]
+ [1]
+
+ $ cat ../server/.hg/blackbox.log
+ *> updated base branch cache in * seconds (glob)
+ *> wrote base branch cache with 1 labels and 1 nodes (glob)
+ *> 1 incoming changes - new heads: 96ee1d7354c4 (glob)
+ *> updated base branch cache in * seconds (glob)
+ *> wrote base branch cache with 1 labels and 1 nodes (glob)
+ *> 1 incoming changes - new heads: 5e849d85a748 (glob)
+
+Activating the hook causes tags cache to get populated
+
+ $ cat >> ../server/.hg/hgrc << EOF
+ > [hooks]
+ > pretxnclose.populate_caches = python:mozhghooks.populate_caches.hook
+ > EOF
+
+ $ hg tag newtag
+ $ hg -q push
+
+ $ [ -f ../server/.hg/cache/hgtagsfnodes1 ]
+
+ $ tail -10 ../server/.hg/blackbox.log
+ *> 1 incoming changes - new heads: 5e849d85a748 (glob)
+ *> updated base branch cache in * seconds (glob)
+ *> wrote base branch cache with 1 labels and 1 nodes (glob)
+ *> writing 72 bytes to cache/hgtagsfnodes1 (glob)
+ *> 0/1 cache hits/lookups in * seconds (glob)
+ *> writing .hg/cache/tags2-served with 2 tags (glob)
+ *> 1/1 cache hits/lookups in * seconds (glob)
+ *> writing .hg/cache/tags2 with 2 tags (glob)
+ *> pythonhook-pretxnclose: mozhghooks.populate_caches.hook finished in * seconds (glob)
+ *> 1 incoming changes - new heads: cf120f74c0ec (glob)
--- a/hgserver/tests/test-push-basic.t
+++ b/hgserver/tests/test-push-basic.t
@@ -60,16 +60,19 @@ Blackbox logging recorded appropriate en
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnopen: hgext_vcsreplicator.pretxnopenhook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-prechangegroup: hgext_readonly.prechangegrouphook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnchangegroup: mozhghooks.single_root.hook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnchangegroup: hgext_pushlog.pretxnchangegrouphook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnchangegroup: hgext_vcsreplicator.pretxnchangegrouphook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> updated base branch cache in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> wrote base branch cache with 1 labels and 1 nodes (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-prepushkey: hgext_readonly.prepushkeyhook finished in * seconds (glob)
+ * user1@example.com @0000000000000000000000000000000000000000 (*)> writing .hg/cache/tags2-served with 0 tags (glob)
+ * user1@example.com @0000000000000000000000000000000000000000 (*)> writing .hg/cache/tags2 with 0 tags (glob)
+ * user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnclose: mozhghooks.populate_caches.hook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pretxnclose: hgext_vcsreplicator.pretxnclosehook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-pushkey: hgext_vcsreplicator.pushkeyhook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-txnclose: hgext_vcsreplicator.txnclosehook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> exthook-changegroup.a_recordlogs: /var/hg/version-control-tools/scripts/record-pushes.sh finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-changegroup: mozhghooks.push_printurls.hook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-changegroup: mozhghooks.advertise_upgrade.hook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> pythonhook-changegroup: hgext_vcsreplicator.changegrouphook finished in * seconds (glob)
* user1@example.com @0000000000000000000000000000000000000000 (*)> 1 incoming changes - new heads: 77538e1ce4be (glob)
